Nursing educators' perspectives on artificial intelligence: A rapid review of roles, limits, and implications.
TL;DR
AI's educational impact depends less on technological capability than on pedagogical design, faculty preparedness, and governance, which indicates that these preconditions are themselves unevenly distributed, highlighting the need for structured implementation strategies that address infrastructure and verification burden alongside pedagogy.
Abstract
Background
Artificial intelligence (AI) is rapidly entering academic nursing education, yet its integration remains uneven and often lacks pedagogical guidance. Understanding how nursing educators perceive AI's role is critical to ensuring its appropriate and effective use.
Aim
This rapid review aimed to synthesize current evidence on nursing educators' perceptions of AI in academic nursing education, with a focus on identifying which educational tasks can be enhanced, replaced, or are not amenable to AI.
Methods
A rapid review was conducted using a multimethod search strategy that combined AI-assisted semantic searching, structured database searches, targeted journal and reference list searching, and manual verification. Eligible studies reporting nursing educators' perspectives on AI in academic settings were synthesized using descriptive and directed content analyses informed by predefined domains, while remaining open to emergent themes.
Results
A total of 55 studies were included in this review. Educators consistently view AI as augmenting rather than replacing faculty roles. AI is perceived as most effective in simulation-based learning and personalized tutoring, followed by feedback and assessment, curriculum design and administrative work, and academic writing and research support. Bounded tasks, including administrative drafting, grading and the summarizing of narrative data, may be partially substituted, though faculty oversight remains necessary. In contrast, relational, ethical and judgment-based domains, including empathy, moral reasoning, complex clinical judgment, hands-on clinical practice and faculty mentorship, are not considered substitutable. Workload effects are directionally mixed: AI reduces time spent on bounded tasks, but verifying its outputs generates new demands. Perceptions vary with prior exposure, which correlated with trust in AI, and with age, gender, academic rank and nationality. Key barriers include limited training, ethical concerns and infrastructure gaps.
Conclusion
AI's educational impact depends less on technological capability than on pedagogical design, faculty preparedness, and governance. Evidence from resource-constrained settings indicates that these preconditions are themselves unevenly distributed, highlighting the need for structured implementation strategies that address infrastructure and verification burden alongside pedagogy.